AI beyond big tech: Why Asia's growth story is just getting started

Key points

• AI spending is broadening beyond U.S. mega-cap technology names into the Asian supply chain.
• Emerging market earnings are being supported by AI-related demand, especially in parts of North Asia.
• Memory, power delivery and cooling are three practical bottlenecks that could shape the next phase of AI investment.
• Selectivity remains important, as supply risks and short-term market swings could affect outcomes.

Investors often associate AI with a handful of large U.S. technology companies. Yet some of the companies enabling the AI buildout sit in Asia, where earnings momentum has been supported by demand for semiconductors, power infrastructure and cooling technologies.

Emerging market (EM) equities are up around 55% over the past year,¹ with earnings a key driver. EM earnings grew roughly 8% in 2025 and are forecast to grow over 35% this year ― more than any other region, as shown below.

Chart: EMs leading global earnings strength

12-month earnings and sales growth estimates by region

Much of this strength comes from the northern part of Asia, with earnings in Taiwan forecast to grow 34% in 2026, and South Korean earnings 220%.²

In South Korea, some of the earnings drivers are particular to the country, such as the government’s corporate reform efforts aimed at increasing shareholder value.

And many South Korean companies, including the shipbuilders, have also benefited from increased geopolitical fragmentation and the desire of some western countries to reduce reliance on China.

Yet the main Asian earnings driver can be summed up in two letters: AI. Much of the enormous AI investment from the big U.S. tech companies ― we estimate up to US$10 trillion by 2030³ ― is flowing to the large Asian makers of the semiconductors essential for AI expansion, as well as the supply chains that have developed around these giants.

How Asia is building and powering AI

For investors, the AI opportunity in Asia can be viewed through three practical bottlenecks: memory, power delivery and cooling.

Advanced AI processes require vast amounts of memory ― for data storage and access ― and vast amounts of power. Shortages of both are supporting the long-term earnings growth for the companies that can help solve these constraints.

Memory demand

Memory shortages are expected to persist for at least two years, with demand around 20%-30% above supply, according to our analysis.⁴ Semiconductor demand is driven by AI training needs and the increasingly complex workloads of AI agents. And it is difficult to quickly expand supply due to the many years it takes to build the new fabrication plants and clean rooms needed to manufacture semiconductors.

Asia is home to some of the world’s leading memory providers, and the top companies are seeking to move up the value chain from simply supplying standardized semiconductors to designing memory to customer specifications. One risk to watch: Over-supply could be a problem in 2027-2028 if demand wanes just as new production facilities come online.

Power demand

AI data center power demand is forecast to quadruple globally over the next decade.⁵ This is a large reason why the International Energy Agency expects global electricity demand to increase by around 4% every year over the next few years, equivalent to adding the electricity consumption of Japan every year.⁶

The investment implication? Power delivery is emerging as one of the next major AI bottlenecks. Several Asian companies are key to the expansion of global power infrastructure, benefiting from surging U.S. gas turbine demand, increasing use of nuclear power and demand for extra high-voltage transformers, which help transfer large amounts of energy to data centers from far-away production sources. Battery technology in China is also providing leading energy storage solutions.

Cooling demand

Cooling is another critical input for energy-intensive data centers, and greater chip power means an increasing shift from air to liquid cooling. Asia is home to leading-edge companies that are adopting next-gen cooling systems that enable greater computing power in the same space.

The bottom line:

We favor businesses that can grow steadily over time, withstanding short-term wobbles, and believe global AI spending has put several Asian companies firmly in this category.